Table 1
Outline of fields, their data types, formats and descriptions.
| FIELD NAME | TYPE | FORMAT/STRUCTURE | DESCRIPTION |
|---|---|---|---|
| record_id | Integer | Numeric | Unique identifier assigned within the internal research database to each auction record. |
| sale_date | String | YYYY-MM-DD | Date on which the auction took place. |
| artist_name_1 | String | Text | Name of the principal artist associated with the auctioned object. |
| object_type | String | Text | Category of object sold (e.g., painting, sculpture, print). |
| auction_house_1 | String | Text | Name of the auction house responsible for conducting the sale. |
| gpi_auction_entry | String | Full text | Complete auction entry text as recorded in the Getty Provenance Index. |
| extracted_text | String | 150-character snippet | Substring automatically extracted from the latter half of the GPI entry, containing references to expert opinions. |
| expert_names | Array of strings | List | Cleaned list of expert surnames identified within the extracted text by the LLM pipeline. |
| titles_in_text | Array of strings | List | Honorifics, academic titles, or institutional designations found in the auction text (e.g., “Dr.”, “Prof.”, “Geh.-Rat”). |
| heidelberg_url | String | URI | Link to the corresponding digitised catalogue page hosted by Heidelberg University Library. |
Figure 1
References to Gutachten and named expert opinions in auction catalogues from 1900–1945.